Upload trainer_state.json
Browse files- trainer_state.json +373 -0
trainer_state.json
ADDED
@@ -0,0 +1,373 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 2.9999490445859873,
|
5 |
+
"global_step": 29436,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.05,
|
12 |
+
"learning_rate": 4.915069982334556e-05,
|
13 |
+
"loss": 1.8335,
|
14 |
+
"step": 500
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.1,
|
18 |
+
"learning_rate": 4.830139964669113e-05,
|
19 |
+
"loss": 1.3358,
|
20 |
+
"step": 1000
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.15,
|
24 |
+
"learning_rate": 4.745209947003669e-05,
|
25 |
+
"loss": 1.2093,
|
26 |
+
"step": 1500
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 0.2,
|
30 |
+
"learning_rate": 4.660279929338225e-05,
|
31 |
+
"loss": 1.1204,
|
32 |
+
"step": 2000
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.25,
|
36 |
+
"learning_rate": 4.5753499116727814e-05,
|
37 |
+
"loss": 1.0652,
|
38 |
+
"step": 2500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 0.31,
|
42 |
+
"learning_rate": 4.490419894007338e-05,
|
43 |
+
"loss": 0.9992,
|
44 |
+
"step": 3000
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.36,
|
48 |
+
"learning_rate": 4.405489876341894e-05,
|
49 |
+
"loss": 0.9696,
|
50 |
+
"step": 3500
|
51 |
+
},
|
52 |
+
{
|
53 |
+
"epoch": 0.41,
|
54 |
+
"learning_rate": 4.32055985867645e-05,
|
55 |
+
"loss": 0.9374,
|
56 |
+
"step": 4000
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 0.46,
|
60 |
+
"learning_rate": 4.235629841011007e-05,
|
61 |
+
"loss": 0.9147,
|
62 |
+
"step": 4500
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 0.51,
|
66 |
+
"learning_rate": 4.150699823345564e-05,
|
67 |
+
"loss": 0.8813,
|
68 |
+
"step": 5000
|
69 |
+
},
|
70 |
+
{
|
71 |
+
"epoch": 0.56,
|
72 |
+
"learning_rate": 4.06576980568012e-05,
|
73 |
+
"loss": 0.8565,
|
74 |
+
"step": 5500
|
75 |
+
},
|
76 |
+
{
|
77 |
+
"epoch": 0.61,
|
78 |
+
"learning_rate": 3.980839788014676e-05,
|
79 |
+
"loss": 0.8328,
|
80 |
+
"step": 6000
|
81 |
+
},
|
82 |
+
{
|
83 |
+
"epoch": 0.66,
|
84 |
+
"learning_rate": 3.895909770349233e-05,
|
85 |
+
"loss": 0.8021,
|
86 |
+
"step": 6500
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.71,
|
90 |
+
"learning_rate": 3.810979752683789e-05,
|
91 |
+
"loss": 0.7917,
|
92 |
+
"step": 7000
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 0.76,
|
96 |
+
"learning_rate": 3.726049735018345e-05,
|
97 |
+
"loss": 0.7935,
|
98 |
+
"step": 7500
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 0.82,
|
102 |
+
"learning_rate": 3.641119717352902e-05,
|
103 |
+
"loss": 0.7608,
|
104 |
+
"step": 8000
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 0.87,
|
108 |
+
"learning_rate": 3.556189699687458e-05,
|
109 |
+
"loss": 0.7619,
|
110 |
+
"step": 8500
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 0.92,
|
114 |
+
"learning_rate": 3.471259682022014e-05,
|
115 |
+
"loss": 0.7359,
|
116 |
+
"step": 9000
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 0.97,
|
120 |
+
"learning_rate": 3.386329664356571e-05,
|
121 |
+
"loss": 0.7315,
|
122 |
+
"step": 9500
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 1.02,
|
126 |
+
"learning_rate": 3.301399646691127e-05,
|
127 |
+
"loss": 0.7218,
|
128 |
+
"step": 10000
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 1.07,
|
132 |
+
"learning_rate": 3.216469629025683e-05,
|
133 |
+
"loss": 0.7217,
|
134 |
+
"step": 10500
|
135 |
+
},
|
136 |
+
{
|
137 |
+
"epoch": 1.12,
|
138 |
+
"learning_rate": 3.131539611360239e-05,
|
139 |
+
"loss": 0.7214,
|
140 |
+
"step": 11000
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 1.17,
|
144 |
+
"learning_rate": 3.0466095936947957e-05,
|
145 |
+
"loss": 0.6976,
|
146 |
+
"step": 11500
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"epoch": 1.22,
|
150 |
+
"learning_rate": 2.961679576029352e-05,
|
151 |
+
"loss": 0.6782,
|
152 |
+
"step": 12000
|
153 |
+
},
|
154 |
+
{
|
155 |
+
"epoch": 1.27,
|
156 |
+
"learning_rate": 2.8767495583639082e-05,
|
157 |
+
"loss": 0.6743,
|
158 |
+
"step": 12500
|
159 |
+
},
|
160 |
+
{
|
161 |
+
"epoch": 1.32,
|
162 |
+
"learning_rate": 2.791819540698465e-05,
|
163 |
+
"loss": 0.6816,
|
164 |
+
"step": 13000
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 1.38,
|
168 |
+
"learning_rate": 2.706889523033021e-05,
|
169 |
+
"loss": 0.6558,
|
170 |
+
"step": 13500
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 1.43,
|
174 |
+
"learning_rate": 2.621959505367577e-05,
|
175 |
+
"loss": 0.6545,
|
176 |
+
"step": 14000
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 1.48,
|
180 |
+
"learning_rate": 2.5370294877021332e-05,
|
181 |
+
"loss": 0.6515,
|
182 |
+
"step": 14500
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 1.53,
|
186 |
+
"learning_rate": 2.4520994700366897e-05,
|
187 |
+
"loss": 0.6412,
|
188 |
+
"step": 15000
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 1.58,
|
192 |
+
"learning_rate": 2.367169452371246e-05,
|
193 |
+
"loss": 0.6214,
|
194 |
+
"step": 15500
|
195 |
+
},
|
196 |
+
{
|
197 |
+
"epoch": 1.63,
|
198 |
+
"learning_rate": 2.2822394347058025e-05,
|
199 |
+
"loss": 0.6273,
|
200 |
+
"step": 16000
|
201 |
+
},
|
202 |
+
{
|
203 |
+
"epoch": 1.68,
|
204 |
+
"learning_rate": 2.197309417040359e-05,
|
205 |
+
"loss": 0.6249,
|
206 |
+
"step": 16500
|
207 |
+
},
|
208 |
+
{
|
209 |
+
"epoch": 1.73,
|
210 |
+
"learning_rate": 2.1123793993749154e-05,
|
211 |
+
"loss": 0.6123,
|
212 |
+
"step": 17000
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 1.78,
|
216 |
+
"learning_rate": 2.0274493817094715e-05,
|
217 |
+
"loss": 0.6152,
|
218 |
+
"step": 17500
|
219 |
+
},
|
220 |
+
{
|
221 |
+
"epoch": 1.83,
|
222 |
+
"learning_rate": 1.942519364044028e-05,
|
223 |
+
"loss": 0.6166,
|
224 |
+
"step": 18000
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"epoch": 1.89,
|
228 |
+
"learning_rate": 1.857589346378584e-05,
|
229 |
+
"loss": 0.6075,
|
230 |
+
"step": 18500
|
231 |
+
},
|
232 |
+
{
|
233 |
+
"epoch": 1.94,
|
234 |
+
"learning_rate": 1.7726593287131404e-05,
|
235 |
+
"loss": 0.5962,
|
236 |
+
"step": 19000
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 1.99,
|
240 |
+
"learning_rate": 1.687729311047697e-05,
|
241 |
+
"loss": 0.5921,
|
242 |
+
"step": 19500
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 2.04,
|
246 |
+
"learning_rate": 1.602799293382253e-05,
|
247 |
+
"loss": 0.5915,
|
248 |
+
"step": 20000
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 2.09,
|
252 |
+
"learning_rate": 1.5178692757168095e-05,
|
253 |
+
"loss": 0.5669,
|
254 |
+
"step": 20500
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 2.14,
|
258 |
+
"learning_rate": 1.4329392580513656e-05,
|
259 |
+
"loss": 0.5725,
|
260 |
+
"step": 21000
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 2.19,
|
264 |
+
"learning_rate": 1.348009240385922e-05,
|
265 |
+
"loss": 0.5728,
|
266 |
+
"step": 21500
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 2.24,
|
270 |
+
"learning_rate": 1.2630792227204785e-05,
|
271 |
+
"loss": 0.5588,
|
272 |
+
"step": 22000
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 2.29,
|
276 |
+
"learning_rate": 1.1781492050550347e-05,
|
277 |
+
"loss": 0.5648,
|
278 |
+
"step": 22500
|
279 |
+
},
|
280 |
+
{
|
281 |
+
"epoch": 2.34,
|
282 |
+
"learning_rate": 1.093219187389591e-05,
|
283 |
+
"loss": 0.5581,
|
284 |
+
"step": 23000
|
285 |
+
},
|
286 |
+
{
|
287 |
+
"epoch": 2.4,
|
288 |
+
"learning_rate": 1.0082891697241474e-05,
|
289 |
+
"loss": 0.5553,
|
290 |
+
"step": 23500
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 2.45,
|
294 |
+
"learning_rate": 9.233591520587037e-06,
|
295 |
+
"loss": 0.5439,
|
296 |
+
"step": 24000
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 2.5,
|
300 |
+
"learning_rate": 8.3842913439326e-06,
|
301 |
+
"loss": 0.5367,
|
302 |
+
"step": 24500
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 2.55,
|
306 |
+
"learning_rate": 7.534991167278163e-06,
|
307 |
+
"loss": 0.5401,
|
308 |
+
"step": 25000
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"epoch": 2.6,
|
312 |
+
"learning_rate": 6.685690990623726e-06,
|
313 |
+
"loss": 0.5302,
|
314 |
+
"step": 25500
|
315 |
+
},
|
316 |
+
{
|
317 |
+
"epoch": 2.65,
|
318 |
+
"learning_rate": 5.83639081396929e-06,
|
319 |
+
"loss": 0.541,
|
320 |
+
"step": 26000
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 2.7,
|
324 |
+
"learning_rate": 4.987090637314853e-06,
|
325 |
+
"loss": 0.5218,
|
326 |
+
"step": 26500
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 2.75,
|
330 |
+
"learning_rate": 4.137790460660416e-06,
|
331 |
+
"loss": 0.5149,
|
332 |
+
"step": 27000
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 2.8,
|
336 |
+
"learning_rate": 3.2884902840059795e-06,
|
337 |
+
"loss": 0.5358,
|
338 |
+
"step": 27500
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 2.85,
|
342 |
+
"learning_rate": 2.4391901073515425e-06,
|
343 |
+
"loss": 0.5353,
|
344 |
+
"step": 28000
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 2.9,
|
348 |
+
"learning_rate": 1.589889930697106e-06,
|
349 |
+
"loss": 0.5171,
|
350 |
+
"step": 28500
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 2.96,
|
354 |
+
"learning_rate": 7.405897540426689e-07,
|
355 |
+
"loss": 0.528,
|
356 |
+
"step": 29000
|
357 |
+
},
|
358 |
+
{
|
359 |
+
"epoch": 3.0,
|
360 |
+
"step": 29436,
|
361 |
+
"total_flos": 2.745512583065764e+17,
|
362 |
+
"train_loss": 0.7166476119453937,
|
363 |
+
"train_runtime": 30706.9194,
|
364 |
+
"train_samples_per_second": 9.586,
|
365 |
+
"train_steps_per_second": 0.959
|
366 |
+
}
|
367 |
+
],
|
368 |
+
"max_steps": 29436,
|
369 |
+
"num_train_epochs": 3,
|
370 |
+
"total_flos": 2.745512583065764e+17,
|
371 |
+
"trial_name": null,
|
372 |
+
"trial_params": null
|
373 |
+
}
|